An exploratory study of manufacturing data and its potential for continuous process improvements from a production economical perspective

Background: Continues improvements in production are essential in order to compete on the market. However, to be an active competitor on the market, companies need to know their strengths and weaknesses, and improve and develop their production continually. Today process industries generate enormous volumes of data and data are considered a valuable source for companies to find new ways to boost their operations' productivity and profitability. Data Mining (DM) is the process of discovering useful patterns and trends in large data sets. Several authors have pointed out data mining as a good data analysis process for manufacturing due to the large amount of data generated and collected from production processes. In manufacturing, DM has two primary goals, descriptive with the focus on discovering patterns to describe the data and predictive where a model is used to determine future values of important variables. Objectives: The objective of this study was to get a deeper understanding of how collected data from production can lead to insights regarding potential production economic improvementsby following the CRISP-DM methodology. In particular to the chosen production line if there were any differences in replenishment durations when it comes to different procedures. Duration in this study is the time the line is halted during a material replenishment. The procedures in question are single-replenishment versus double-replenishment. Further investigated was if there were any differences in the replenishment duration when it comes to which shift team and at what shift time the replenishment procedures were made. Methods: In this study the CRISP-DM methodology was used for structuring the collected data from the case company. The data was primarily historical data from a continues production process. To verify the objective of the study, three hypotheses derived from the objective was tested by using a t test and Bonferroni test. Results: The result showed that the duration of a double-replenishment is lower compared to two single-replenishments. Further results showed that there is a significant difference in the single-replenishment duration between the different shift times and different working teams. The interpretation of the result is that in the short term there is a possibility that implementingdouble replenishments can reduce the throughput time and possibility also the lead time. Conclusions: This study could contribute with knowledge for others who seek a way to use data to detect information or deeper knowledge about a continuous production process. The findings in this study could be specifically interesting for cable manufacturers and, in general, for continuous process manufacturers. Further conclusions are that time-based competition is one way for increasing the competitive advantage in the market. By using manufacturing generated data, it is possible to analyse and find valuable information that can contribute to continuous process improvements and increase the competitive advantage